自然语言处理的新需求、新资源、新技术.ppt

《自然语言处理的新需求、新资源、新技术.ppt》由会员分享，可在线阅读，更多相关《自然语言处理的新需求、新资源、新技术.ppt（19页珍藏版）》请在三一办公上搜索。

1、自然语言处理的新需求、新资源、新技术,董振东 董强Tel:(8610)6287-5641 Tel:(8610)6676-8816 Intel China Research Forum北京 2000/10/11,提纲差距与反思-成就是过去时，不足是将来时知网-一种新的资源-新资源促进新技术,差距与反思,层次浅语境窄知识颗粒度小流行技术的极限,音字转换信息过滤文本分析语音识别文本分类歧义排除OCR信息检索文本理解语法检查自动文摘机器翻译,音字转换系统（输入法）上这摊某被立即送往医院，但终因史学过多，不止身亡。伤者谭某被立即送往医院，但终因失血过多，不治身亡。上这里某被立即送往医院，但终因留学过多，

2、不止身亡。伤者李某被立即送往医院，但终因流血过多，不治身亡。唐非情此获准。唐飞请辞获准。,语法检查In a few years time,there will be no Internet companies-there will just be companies-and all companies that are going to operate in the economics of a few years,in the future,are going to be Internet companies.-Andrew S.Grove(1)areThe world will littl

3、e note,nor long remember what we say here,but it can never forget what they did here.-Abraham Lincon(1)world wills/worlds will(2)will little neither note,nor/will no little note,orIt would be something that we actually will almost take for granted and wonder how business was done before we incorpora

4、ted this in a very deep way-Bill Gates(1)wonders,信息检索华人 中华人民共和国，新华人寿保险北大 北大西洋，北大荒，台北大学葡萄 葡萄牙，葡萄糖,自动文摘Advances in Automatic Text SummarizationWith the rapid growth of the World Wide Web and electronic information services,information is becoming available on-line at an incredible rate.One result is t

5、he oft-decried information overload.No one has time to read everything,yet we often have to make critical decisions based on what we are able to assimilate.The technology of automatic text summarization is becoming indispensable for dealing with this problem.Text summarization is the process of dist

6、illing the most important information from a source to produce an abridged version for a particular user or task.Until now there has been no state-of-the-art collection of the most important writings in automatic text summarization.This book presents the key developments in the field in an integrate

7、d framework and suggests future research areas.The book is organized into six sections:Classical Approaches,Corpus-Based Approaches,Exploiting Discourse Structure,Knowledge-Rich Approaches,Evaluation Methods,and New Summarization Problem Areas.（55%）,自动翻译Do you think I could stay here and become noth

8、ing to you?Do you think,because I am poor,and obscure and plain,that I am soulless and heartless?I have as much soul as you and fully as much heart.And if God had gifted me with wealth and beauty I should have made it as hard for you to leave me as it is now for me to leave you.There I have spoken m

9、y heart and let me go.你想我可以在这里停留并且适合没有任何东西成为你吗？你想，因为我贫穷，并且昏暗和简明，我是卑鄙和无情的吗？我有和灵魂AS多你和充分作为许多心脏。同时如果上帝有我天才财富以及美丽我作为应该使它努力为你因为它现在为我。我已在那里说我的心脏并且让我去。(1988年推出的系统)你认为我能这里留下和变成对你没有什么吗?你因为我是可怜和不引人注目和清楚,认为我是没有灵魂的和无情吗?我有同样多作为你精神和完全同样多心脏.和如果上帝有有天赋有财富的我和美,我应该已使它变得你同样地难以现在我留下你让我保持现在的样子.那里我说我的心脏已和让我去。(1999年推出的系统),

10、知网-一种新的资源,知网的近期发展与应用知网的关键,HowNet 意味着什么？How knowledge is represented and acquired?How meaning can be formalized and calculated?How meaning is expressed and conveyed?,知网近期的发展,知网2000版与1999版的不同2000版1999版语种中（GB）-英双语 GB+Big5功能浏览器+数据维护基本数据知识词典11万记录6万记录新增修订10,000余增加中文多义义项例子内容扩展事件关系与角色转换库知网-中文信息结构库,知网近期的应用,资

11、源扩展语义标注建立关系网络信息处理应用语义分析排除歧义英中-中英双向机器翻译信息过滤,知网的关键,知网的灵魂关系-关系的动态的、多层次的体现寓于静态的、孤立的描述之中意义的形式化、可计算知网应用的关键-新技术的引入大语境 可能且受鼓励元规则的机制,举例：,我上星期把自行车卖了。今天一大早买主来找我，他说那车的车身有过硬伤，他要退货。买主-the buyer车 bycicle(bike)?or car?车身-the body of,sell|卖agent,possession,target,costbuy|买(X)sell|卖(Y)mutual implication;agent OF buy|

12、买=target OF sell|卖;source OF buy|买=agent OF sell|卖;possession OF buy|买=possession OF sell|卖;cost OF buy|买=cost OF sell|卖.W_C=买主G_C=NE_C=W_E=buyerG_E=NE_E=DEF=human|人,#commercial|商,*buy|买,W_C=自行车G_C=NE_C=W_E=bicycleG_E=NE_E=DEF=LandVehicle|车车 last查询,W_C=车身G_C=NE_C=W_E=body of a vehicleG_E=NE_E=DEF=pa

13、rt|部件,%LandVehicle|车,body|身,唐飞请辞获准。SYN_S=V-VSEM_S=(事件,行动,使动/阻动)-结果事件(事件)请-示，请-转，请-来，请-教，请-吃，请-喝，请-提意见，SYN_S=V-VSEM_S=(事件)递续-(事件)举枪-射击，拔枪-射击，拔刀-相助，拜师-学艺，打击-报复，讨论-决定，讨论-通过，立案-侦查，报到-上班，挂号-交费，细嚼-慢咽，登台-献艺，握手-告别，前来-报到，列队-欢迎，泛滥-成灾，拨号-接通，离家-出走，纵火-焚烧，改恶-从善，走私-贩私，出席-作陪，防火-护林，封山-育林，退耕-还林，抗洪-救灾，团结-互助，团结-奋进，停职-检查，不治-身亡，,知网的未来,-共建、共享扩充词语量、语种应用中改进,欢迎提问和批评,